Spent 2 years prompting AI wrong until a coworker said 5 words
I was writing these long, detailed prompts for our customer service bot and getting flat answers every time. A coworker glanced over and said, "you're giving it a novel, not a task." Shortened everything to clear steps with one example each and the quality jumped overnight. Has anyone else found their prompts improve when they cut the extra words?
And honestly that same thing shows up everywhere once you start looking. I've noticed with recipes, if someone gives me a paragraph about their grandmother's garden before the actual ingredients, I just skim it and miss stuff. Same with work emails, people bury the actual question under three paragraphs of context and then wonder why nobody answers. It's like our brains have a limit on how much we can hold at once, so extra words just become noise. The real skill is knowing what to leave out, not what to add. Once you trim the fluff, the core instruction finally gets room to land.